Search in sources :

Example 6 with Response

use of org.jsoup.Connection.Response in project ripme by RipMeApp.

the class SankakuComplexRipper method getFirstPage.

@Override
public Document getFirstPage() throws IOException {
    if (albumDoc == null) {
        Response resp = Http.url(url).response();
        cookies.putAll(resp.cookies());
        albumDoc = resp.parse();
    }
    return albumDoc;
}
Also used : Response(org.jsoup.Connection.Response)

Example 7 with Response

use of org.jsoup.Connection.Response in project crawler-jsoup-maven by bluetata.

the class ITEYELoginApater method login.

/**
 * 模拟登陆Iteye
 * @param userName
 * @param pwd
 * @throws Exception
 */
public static void login(String userName, String pwd) throws Exception {
    // 第一次请求
    // 获取连接
    Connection con = Jsoup.connect("http://www.iteye.com/login");
    // 配置模拟浏览器
    con.header("User-Agent", "Mozilla/5.0 (Windows NT 6.1; WOW64; rv:29.0) Gecko/20100101 Firefox/29.0");
    // 获取响应
    Response rs = con.execute();
    // 转换为Dom树
    Document d1 = Jsoup.parse(rs.body());
    // 获取form表单,可以通过查看页面源码代码得知
    List<Element> et = d1.select("#login_form");
    // 获取,cooking和表单属性,下面map存放post时的数据
    Map<String, String> datas = new HashMap<>();
    for (Element e : et.get(0).getAllElements()) {
        if (e.attr("name").equals("name")) {
            // 设置用户名
            e.attr("value", userName);
        }
        if (e.attr("name").equals("password")) {
            // 设置用户密码
            e.attr("value", pwd);
        }
        if (e.attr("name").length() > 0) {
            // 排除空值表单属性
            datas.put(e.attr("name"), e.attr("value"));
        }
    }
    /**
     * 第二次请求,post表单数据,以及cookie信息
     */
    Connection con2 = Jsoup.connect("http://www.iteye.com/login");
    con2.header("User-Agent", "Mozilla/5.0 (Windows NT 6.1; WOW64; rv:29.0) Gecko/20100101 Firefox/29.0");
    // 设置cookie和post上面的map数据
    Response login = con2.ignoreContentType(true).method(Method.POST).data(datas).cookies(rs.cookies()).execute();
    // 打印,登陆成功后的信息
    System.out.println(login.body());
    // 登陆成功后的cookie信息,可以保存到本地,以后登陆时,只需一次登陆即可
    Map<String, String> map = login.cookies();
    for (String s : map.keySet()) {
        System.out.println(s + " : " + map.get(s));
    }
}
Also used : Response(org.jsoup.Connection.Response) HashMap(java.util.HashMap) Element(org.jsoup.nodes.Element) Connection(org.jsoup.Connection) Document(org.jsoup.nodes.Document)

Example 8 with Response

use of org.jsoup.Connection.Response in project crawler-jsoup-maven by bluetata.

the class AmazonLoginApater method main.

public static void main(String[] args) {
    // grab login form page first
    try {
        // lets make data map containing all the parameters and its values found in the form
        Map<String, String> mapParamsData = new HashMap<String, String>();
        mapParamsData.put("email", "dietime1943@hotmail.com");
        mapParamsData.put("password", "152300");
        Response loginResponse = Jsoup.connect("https://passport.jd.com/new/login.aspx").userAgent(USER_AGENT).timeout(TIMEOUT_UNIT * TIMEOUT_TIMES).data(mapParamsData).method(Method.POST).followRedirects(true).execute();
        System.out.println("Fetched login page");
        // System.out.println(loginResponse.toString());
        // get the cookies from the response, which we will post to the action URL
        Map<String, String> mapLoginPageCookies = loginResponse.cookies();
        System.out.println(mapLoginPageCookies);
    } catch (Exception e) {
        e.printStackTrace();
    }
}
Also used : Response(org.jsoup.Connection.Response) HashMap(java.util.HashMap)

Example 9 with Response

use of org.jsoup.Connection.Response in project crawler-jsoup-maven by bluetata.

the class TestApiOfConnect method main.

public static void main(String[] args) throws IOException {
    // TODO Auto-generated method stub
    // Connection connection = Jsoup.connect("http://bluetata.com");
    // // connection.data("aaa","ccc");  // 这是重点
    // 
    // connection.header("Content-Type", "application/json; charset=UTF-8");  // 这是重点
    // 
    // connection.header("Accept", "text/plain, */*; q=0.01");
    // 
    // connection.timeout(15000);
    // 
    // //String body = "{\"CategoryType\":\"SiteHome\",\"ParentCategoryId\":0,\"CategoryId\":808,\"PageIndex\":2,\"TotalPostCount\":4000,\"ItemListActionName\":\"PostList\"}";
    // 
    // //connection.requestBody(body);
    // 
    // Document document = connection.post();
    String jsonBody = "{\"name\":\"ACTIVATE\",\"value\":\"E0010\"}";
    Connection connection = Jsoup.connect("http://bluetata.com/").userAgent(// User-Agent of Chrome 55
    "Mozilla/5.0 (Windows NT 6.3; Win64; x64) AppleWebKit/537.36 (KHTML, like Gecko) Chrome/55.0.2883.87 Safari/537.36").referrer("http://bluetata.com/").header("Content-Type", "application/json; charset=UTF-8").header("Accept", "text/plain, */*; q=0.01").header("Accept-Encoding", "gzip,deflate,sdch").header("Accept-Language", "es-ES,es;q=0.8").header("Connection", "keep-alive").header("X-Requested-With", "XMLHttpRequest").requestBody(jsonBody).maxBodySize(100).timeout(1000 * 10).method(Connection.Method.POST);
    Response response = connection.execute();
}
Also used : Response(org.jsoup.Connection.Response) Connection(org.jsoup.Connection)

Example 10 with Response

use of org.jsoup.Connection.Response in project portfolio by buchen.

the class YahooFinanceQuoteFeed method loadCrump.

private Crumb loadCrump(String tickerSymbol) throws IOException {
    // $NON-NLS-1$
    String url = MessageFormat.format("https://de.finance.yahoo.com/quote/{0}/history?p={0}", tickerSymbol);
    Response response = // 
    Jsoup.connect(url).userAgent(OnlineHelper.getUserAgent()).header("Accept", // $NON-NLS-1$ //$NON-NLS-2$
    "text/html,application/xhtml+xml,application/xml;q=0.9,*/*;q=0.8").header("Accept-Language", // //$NON-NLS-1$ //$NON-NLS-2$
    "en-US,en;q=0.5").timeout(// 
    30000).execute();
    // $NON-NLS-1$
    String KEY = "\"CrumbStore\":{\"crumb\":\"";
    String body = response.body();
    int startIndex = body.indexOf(KEY);
    if (startIndex < 0)
        throw new IOException(Messages.MsgErrorNoCrumbFound);
    int endIndex = body.indexOf('"', startIndex + KEY.length());
    if (endIndex < 0)
        throw new IOException(Messages.MsgErrorNoCrumbFound);
    String crumb = body.substring(startIndex + KEY.length(), endIndex);
    crumb = StringEscapeUtils.unescapeJava(crumb);
    return new Crumb(crumb, response.cookies());
}
Also used : Response(org.jsoup.Connection.Response) IOException(java.io.IOException)

Aggregations

Response (org.jsoup.Connection.Response)28 Document (org.jsoup.nodes.Document)12 IOException (java.io.IOException)11 Element (org.jsoup.nodes.Element)10 HashMap (java.util.HashMap)8 URL (java.net.URL)4 Connection (org.jsoup.Connection)4 Elements (org.jsoup.select.Elements)4 ArrayList (java.util.ArrayList)3 BufferedWriter (java.io.BufferedWriter)1 File (java.io.File)1 FileOutputStream (java.io.FileOutputStream)1 FileWriter (java.io.FileWriter)1 HttpURLConnection (java.net.HttpURLConnection)1 MalformedURLException (java.net.MalformedURLException)1 ParseException (java.text.ParseException)1 LocalDate (java.time.LocalDate)1 DateTimeParseException (java.time.format.DateTimeParseException)1 Map (java.util.Map)1 Matcher (java.util.regex.Matcher)1